Can I trust AI on food safety?
The short answer: Not safely
The explanation: Artificial intelligence (AI) systems, such as ChatGPT, Perplexity, and Google AI Overview, gather information from across the Internet to respond to questions. This means their data includes personal opinion from social media and non-science-based websites as well as from peer-reviewed, validated research from scientific literature. When combined, there can a vast range of error and accuracy in any given response.
In fact, an AI statement on the accuracy of AI states that while it is rapidly growing, “it is not yet 100% reliable as a sole, independent safeguard.” And even a small amount of inaccuracy in food safety can be a highly dangerous thing.
It also is important to understand that AI is only as good as the data on which it was trained and often doesn’t have current information; it can provide what seems to be a legitimate explanation for a false statement; and it will even make up information itself (such as citing nonexistent websites). Probably of most concern are the near misses: the information about a food risk that is close to correct, but just enough off so as to be of danger.
So, what can you do to ensure you’re getting accurate information?
If you choose to use AI as a source:
- Always check its source. If a link is provided, check it out to ensure it is credible and correctly interpreted by AI. If there is no link or source provided, it’s better to find another source.
- Use AI as a basis for a general understanding of a topic, then dig deeper using credible sources. Don’t use it as your only reference.
- Ask the question in multiple ways and/or through various AI systems. If the answer changes, it is imperative to seek out a non-AI source.
While AI can provide general information on topics, it is risky to use it as a key source for health or safety advice. Better sources of information are the CDC, FDA, USDA, or university websites.
TAG can simplify your food safety and quality practices. Contact us for information!
